The latest articles on DEV Community by Gregory (@gregory_868542352c4395a2b). https://dev.to/gregory_868542352c4395a2b https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4114236%2F09d335ff-d367-4d33-ab03-b71fa61444cc.png https://dev.to/gregory_868542352c4395a2b en Gregory Mon, 07 Sep 2026 15:46:26 +0000 https://dev.to/gregory_868542352c4395a2b/what-i-learned-building-a-3d-reading-game-my-6-year-old-actually-finishes-3ddo https://dev.to/gregory_868542352c4395a2b/what-i-learned-building-a-3d-reading-game-my-6-year-old-actually-finishes-3ddo <p>Two years ago I stopped buying phonics flashcards and started writing browser games instead. Not because I think software beats a parent reading aloud — it does not — but because the twenty minutes between dinner and bath time were being lost to whatever autoplayed next. If a screen was going to win those minutes, I wanted the screen to teach something.</p> <h2> The first version failed for a boring reason </h2> <p>My first build was a word-matching game: pick the picture that matches the word. It was correct, it was fast, and my daughter quit after four minutes. The problem was not difficulty. The problem was that nothing happened when she was right. A green tick is not a reward; it is a receipt.</p> <p>The rewrite put the answer inside the action. Solve the word and the catapult fires. Miss it and the shot goes wide, nothing explodes, and you try again. Same content, completely different session length: four minutes became twenty-five. That rule — <strong>the shot must be the solved task, not a screen that appears after it</strong> — has survived every game we have built since.</p> <h2> Three things that turned out to matter more than graphics </h2> <p><strong>Teach before you test.</strong> Every new mechanic gets a short, playable demonstration before the first scored round. Kids who are shown once will try; kids who are dropped into a quiz guess and then quit.</p> <p><strong>Errors have to be soft.</strong> No sirens, no red flash, no life counter draining. A miss costs a shot and nothing else. Anxiety and phonics do not mix — a child sounding out "th" needs room to be wrong out loud.</p> <p><strong>Read everything aloud.</strong> Full voice-over on every prompt is the single most expensive feature we ship and the one that decides whether a five-year-old can play alone. A reading game that requires reading to understand the instructions is a locked door.</p> <h2> Where it ended up </h2> <p>The games run in the browser on WebGL, no install, no account, no ads. If you want to see how the pieces fit together, the reading set is at <a href="https://dad4kids.com/en/reading-games-for-kids/" rel="noopener noreferrer">reading games for kids</a> — letters, phonics, sight words, each one built around the same "solve it to act" loop. There is a <a href="https://dad4kids.com/en/math-games-for-kids/" rel="noopener noreferrer">math set</a> built the same way, and a <a href="https://dad4kids.com/en/games-for-preschoolers/" rel="noopener noreferrer">collection for preschoolers</a> where the reading load is near zero.</p> <h2> The honest caveat </h2> <p>None of this replaces books, and screen time is still screen time. What changed in our house is what the minutes are made of. Twenty minutes of a child sounding out words to launch something is not the same twenty minutes as passive video, even though the parental-control app counts them identically. That gap is the whole reason I keep building.</p> frontend gamedev learning programming